Cholesterol- and ssDNA-binding fusion protein-mediated DNA tethering on the plasma membrane

Abstract

DNA modification of the plasma membrane is an excellent approach for controlling membrane–protein interactions, modulating cell–cell/cell–biomolecule interactions, and extending the biosensing field. The hydrophobic insertion of DNA conjugated with hydrophobic anchoring molecules is utilized for tethering DNA on the cell membrane. In this study, we developed an alternative approach to tether DNA on the plasma membrane based on ssDNA- and cholesterol-binding proteins. We designed a fusion protein (Rep–ALOD4) composed of domain 4 of anthrolysin O (ALOD4), which binds to cholesterol in the plasma membrane, and a replication initiator protein derived from porcine circovirus type 2 (Rep), which forms covalent bonds with single-stranded DNA (ssDNA) with a Rep recognition sequence. Rep–ALOD4 conjugates ssDNA to Rep and binds to the plasma membrane via cholesterol, thus tethering ssDNA to the cells. Quartz crystal microbalance measurements showed that membrane cholesterol binding of Rep–ALOD4 to the lipid bilayer containing cholesterol was accelerated above 20% (w/w) cholesterol in the lipid bilayer. Rep–ALOD4 was conjugated to fluorescein-labeled ssDNA (S-FITC–Rep–ALOD4) and used to treat human cervical tumor HeLa cells. The green signal assigned to S-FITC–Rep–ALOD4 was detected along HeLa cells, whereas diminished by cholesterol removal with methyl β-cyclodextrins. Moreover, ssDNA-conjugated Rep–ALOD4 tethered ssDNA-conjugated functional proteins on the HeLa cell plasma membrane via complementary base pairing. Collectively, Rep–ALOD4 has the potential as an ssDNA-tethering material via plasma membrane cholesterol to extend cell surface engineering.

---

Introduction

The plasma membrane plays a central role in regulating cell functions such as cell adhesion, cell–cell recognition, endocytosis, and signal transduction.^1–3 Surface modification of the plasma membrane with exogenous biomaterials such as DNA, RNA, peptides, proteins, and synthetic polymers is an approach for regulating cellular functions. Among them, modification with DNAs provides a unique DNA-derived function for the plasma membrane surface.^4–6 Framework nucleic acids, modified on the plasma membranes, enhance the chemical and biological properties of cells based on their complementary pairing, steric hindrance, molecular recognition, and enzymatic function of DNA folding.^7–11 The modification of the plasma membrane with DNAs has practical applications in therapeutics, tissue engineering, and biosensing.^12–14

Several approaches for the chemical or physicochemical modification of plasma membranes with DNAs have been developed. In chemical modification approaches, functional groups on plasma membrane components, such as amines, thiols, carboxyls, diols, and azides, are used to form chemical covalent bonds with DNAs modified with reactive functional groups.¹⁵ The function of plasma membrane components, including the membrane proteins and glycocalyx, can be potentially impaired by reacting with reactive functional groups. In contrast, in physicochemical modification approaches, DNAs are anchored on plasma membranes via hydrophobic moieties, such as lipids, vitamin E, and cholesterols, which are conjugated to the DNA terminus.¹⁶ Xiong et al. developed aptamer-modified immune cells for cell-based therapy using the designed DNA aptamer conjugated with diacyllipid composed of two stearic acids via poly(ethylene glycol) linkers.¹⁷ In this regard, the hydrophobicity of hydrophobic moieties conjugating to DNA affected the structure of modified DNA and its accessibility to plasma membranes.^18–21 The position and number of hydrophobic moieties conjugating to DNA regulated the mode of DNA insertion into plasma membranes, achieving the development of DNA-based nanopores.^18,22 Additionally, the PEG linker between DNA and hydrophobic moieties reduced the internalization of modified DNA via endocytosis and a nonspecific interaction between the plasma membrane and DNA.^23,24 Collectively, the efficient modification with DNA to the plasma membrane requires various chemical modifications and precise design. The development of alternative methods for DNA modification on plasma membranes without chemical modification and hydrophobic moieties is necessary.

Domain 4 of anthrolysin O (ALOD4, molecular weight: 15 000) is a cholesterol-binding protein.^25–27 Anthrolysin O is a pore-forming bacterial toxin derived from Bacillus anthracis that induces the lysis of red blood cells; ALOD4 that lacked anthrolysin O sub-domains could bind to the plasma membrane surface via cholesterol without any damage.^28,29 Although ALOD4 may be a potential candidate as a DNA-tethering material on plasma membranes via cholesterol, modifying ALOD4 with DNAs requires chemical and biochemical conjugation, leading to multi-step manipulation. In this regard, we have developed a replication initiator protein derived from a bacteriophage (A*, molecular weight: 38 700) fused with proteins of interest (POIs) that added a single-stranded DNA (ssDNA)-binding ability to the POIs.³⁰ Recently, Lovendahl et al. have reported that another replication initiator protein derived from porcine circovirus type 2 (Rep, molecular weight: 12 000) can be used instead of A* because of the small molecular weight and its high activity.³¹ Rep could recognize an ssDNA with a Rep-recognizing sequence (AAGTATTAC) at the 5′ terminus, which formed a covalent bond between Tyr96 of Rep and 5′ terminus ssDNA. We have synthesized Rep-fused POIs, such as bioluminescence proteins, fluorescence proteins, and elastin-like polypeptides, for constructing DNA aptamer-based sandwich assay systems, bioluminescence resonance energy transfer-based biosensors via DNA aptamer–protein hybrid molecules, and ssDNA-conjugated protein nanoparticles to deliver biomolecules and detect tumor cells.^32–35 Hence, Rep fusion with ALOD4 would authorize ALOD4 to bind to ssDNA without chemical modification based on the ssDNA-binding ability of Rep. ssDNA-conjugated Rep–ALOD4 could tether ssDNA on the plasma membrane via ALOD4 binding to cholesterol.

In this study, we developed the DNA modification method without chemical modification and hydrophobic moieties by using the fusion proteins of Rep fused with ALOD4 via a (GGGS)₂ linker to conjugate ssDNA on the plasma membrane via inherent cholesterol (Fig. 1). Rep–ALOD4 bound to liposome-coated substrates containing >20% (w/w) cholesterol via the ALOD4 domain and reacted with an ssDNA containing Rep-recognizing sequence at 37 °C within 30 min via the Rep domain. ssDNA-conjugated Rep–ALOD4 (ssDNA–Rep–ALOD4) bound to the plasma membrane of human cervical carcinoma (HeLa) cells, and ssDNA was bound on the plasma membrane via cholesterol. In addition, a fusion protein composed of the Rep N-terminus fused to the fluorescent protein Venus (Venus–Rep) was constructed as a model functional protein. Venus–Rep and Rep–ALOD4 were bound to the complementary ssDNA using Rep. Venus–Rep was displayed on the plasma membrane via complementary base pairing between ssDNA–Rep–ALOD4 and ssDNA-conjugated Venus–Rep. Taken together, Rep–ALOD4 has the potential as an ssDNA-tethering material to the cell surface via plasma membrane cholesterol.

Fig. 1 Illustration for Rep–ALOD4 showing ssDNA on the plasma membrane of cells via cholesterol.

Materials and methods

Materials

The enzymes used for plasmid construction were purchased from Takara Bio (Shiga, Japan). Escherichia coli KRX strain for protein expression was obtained from Promega (Madison, WI, USA). The bicinchoninic acid (BCA) protein assay kit and SnakeSkin Dialysis Tubing (molecular weight cut-off (MWCO): 10 000 and 30 000, respectively) were purchased from Thermo Fisher Scientific (Grand Island, NY, USA). Lysozyme, isopropyl-β-D(−)-thiogalactopyranoside (IPTG), and imidazole were obtained from FUJIFILM Wako Pure Chemicals (Osaka, Japan). The Profinity IMAC resin was obtained from Bio-Rad. 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC) and cholesterol were obtained from Tokyo Chemical Industry (Tokyo, Japan). ssDNAs with AAGTATTAC at 5′-terminus were synthesized by FASMAC (Kanagawa, Japan) (Table 1). All other solvents and reagents were purchased from FUJIFILM Wako Pure Chemicals and Nacalai Tesque (Kyoto, Japan).

Table 1 ssDNA sequence used in this study

Code	ssDNA sequence^a
a AAGTATTAC, shown in italic type, indicates a sequence recognized by Rep.
S-1	AAGTATTACCAGCGCACTTCGGCAG
S-FITC	AAGTATTACTTTTTTTTTTTTTTTTTTTT-FITC
S-pT	AAGTATTACTTTTTTTTTTTTTTTTTTTT
S-pA	AAGTATTACAAAAAAAAAAAAAAAAAAAA

---

Rep–ALOD4-encoding plasmid DNA construction

Plasmid DNAs encoding Rep–ALOD4 were constructed using the plasmid DNAs previously constructed in our laboratory.^32–35 Briefly, the double-digested gene fragment of (GGGS)₂–ALOD4 with BamHI and XhoI restriction sites at the 5′ and 3′ terminus, respectively, was prepared from pEX–A2J2–(GGGS)₂–ALOD4. The fragment was cloned into pET–Rep–histidine₆ digested with BamHI and XhoI to obtain pET–Rep–(GGGS)₂–Rep–ALOD4, which encodes Rep–ALOD4. E. coli KRX strain was transformed with the constructed plasmids and the DNA sequence was analyzed to verify the correct sequence and corresponding amino acid sequences. The amino acid sequences of the fusion proteins are described in ESI Fig. S1.† Moreover, the hexapeptide sequence of ALOD4 (GTTLYP; amino acid 98–103), which is related to the cholesterol-binding sites, was substituted with AAAAAA via point mutagenesis to obtain modified Rep–ALOD4.²⁷

Protein expression and purification

E. coli KRX strain transformed with the Rep–ALOD4-encoding plasmids were spread on lysogeny broth (LB) agar plates supplemented with 20 μg mL⁻¹ kanamycin and incubated for 18 h at 37 °C. A single colony was inoculated into 5 mL LB medium supplemented with 20 μg mL⁻¹ kanamycin and incubated for 20 h at 37 °C with shaking at 200 rpm. These cultures were transferred to 800 mL LB medium supplemented with 20 μg mL⁻¹ kanamycin, and incubated at 37 °C with shaking at 200 rpm until O.D.600 = 0.4. Then, (0.1%) rhamnose was added to the cultures and incubated at 25 °C with shaking at 125 rpm for 20 h. The transformed E. coli KRXs were collected by centrifugation at 5000g. The collected cells were suspended in 50 mM tris(hydroxymethyl)aminomethane (Tris)-HCl buffer (pH 8.0) containing 150 mM NaCl and 50 μg mL⁻¹ lysozyme, and lysed with freeze–thaw cycles and sonication. Rep–ALOD4 was purified using the Profinity IMAC resin. Purified protein solutions were dialyzed against phosphate-buffered saline (PBS) for 36 h. Protein solutions were stored at 4 °C for further use. The protein solution concentrations were determined using a BCA assay kit. The purity and expected molecular mass of the proteins were confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie Brilliant Blue (CBB) staining.

ssDNA binding to Rep–ALOD4

Briefly, 10 μM Rep–ALOD4 was incubated with 1–10 μM S-1 ssDNA in the presence of Mg²⁺ at 37 °C for arbitrary time. After the reaction, the solution was purified using Amicon Ultra-0.5 centrifugal filter units (molecular weight cutoff: 30 000) (Millipore). The proteins were analyzed by SDS-PAGE and stained with CBB. The Rep–ALOD4 reactivity with ssDNA was calculated using CBB-stained band areas assigned to unmodified Rep–ALOD4.

Liposome preparation

DOPC and cholesterol (10 mg mL⁻¹ total lipid) were dissolved in chloroform at weight ratios of 10 : 0, 9 : 1, 8 : 2, 7 : 3, and 6 : 4. The organic solvent was evaporated under a nitrogen flow and the lipids were maintained under vacuum for 24 h to completely remove the solvent. The lipid film was then hydrated in 1 mL of 50 mM Tris buffer (pH 7.5) containing 150 mM NaCl at 25 °C for 30 min. The lipid suspension was extruded 10 times through polycarbonate membranes (100 nm pore size, Cytiva, MA, USA) at 25 °C using a mini-extruder (Avanti Polar Lipids, Alabaster, AL).

Quartz crystal microbalance (QCM)

The QCM was monitored at 25 °C using a QCM922A system (Seiko EG&G, Tokyo, Japan). Lipid bilayer-coated sensor chip surfaces were prepared on a SiO₂ sensor (QA-A9M-SIO₂-S(SEP), Seiko, Tokyo, Japan) by injecting liposomes for 1 h at 50 μL min⁻¹ using a peristaltic pump (MP-4001, Tokyo Rikakikai Co., Tokyo, Japan). After washing with Tris buffer, Rep–ALOD4 dissolved in Tris buffer was injected for 60 min at 50 μL min⁻¹ at 25 °C. The sensor chip surfaces were then rinsed with Tris buffer for 20 min at 50 μL min⁻¹. The frequency change (ΔF) was recorded to calculate the amount of adsorbed Rep–ALOD4 on the sensor chip using the Sauerbrey equation:
where F₀, A, ρ_q, and μ_q are the fundamental frequency of the quartz crystal (8.9 MHz), the gold sputtered electrode area (0.196 cm²), the quartz density (2.65 g cm⁻³), and the shear modulus of quartz (2.95 × 10⁹ g cm⁻¹ s⁻²), respectively.

Cell culture

Human cervical carcinoma HeLa and human normal diploid fibroblast TIG-3 cells were obtained from the RIKEN BioResource Research Center. Human ovarian cancer SKOV-3 cells, human breast cancer MCF-7 cells, and mouse macrophage-like RAW264.7 cells were provided by American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Wako, Osaka, Japan) supplemented with 10% (v/v) fetal bovine serum (FBS, Biowest) and 1% penicillin–streptomycin at 37 °C and 5% CO₂.

Confocal laser scanning microscopy (CLSM) and flow cytometry

Cells cultured in 100 mm dishes were collected by trypsinization. In a 1.5 mL microcentrifuge tube, 3 × 10⁴ cells were dispersed in 1 mL serum-free medium containing 1, 3, or 5 μM S-FITC–Rep–ALOD4 and incubated for 1 h at 37 °C with rotation. Then, the cells were washed twice with PBS at 37 or 4 °C to remove the unbound proteins. For CLSM measurement, the cells resuspended in 10% FBS-containing DMEM were added to an unmodified glass-bottom dish (3.5 cm, IWAKI). The fluorescence signals of S-FITC–Rep–ALOD4-treated cells were detected using confocal laser scanning microscopy (CLSM; FV-300; Olympus, Japan). For acquiring Z-stack images, HeLa cells were adhered on a glass-bottom dish for 24 h, and treated with 3 μM S-FITC–Rep–ALOD4 for 10 min. Z-Stack images were acquired with Zeiss LSM780 (Zeiss, Germany) equipped with Plan-Apochromat 100×/1.46 oil objective lens. For flow cytometry analysis, the suspended cells were transferred to PBS containing 1% BSA and analyzed using a flow cytometer (RF-500; Sysmex Corporation, Kobe, Japan).

Venus–Rep binding on the plasma membrane

Rep–ALOD4 and Venus–Rep reacted with S-pT (S-pT–Rep–ALOD4) and S-pA (Venus–Rep–S-pA). HeLa cells cultured in 100 mm dishes were washed three times with PBS, followed by trypsinization. In a 1.5 mL microcentrifuge tube, 3 × 10⁴ cells were dispersed in 1 mL serum-free medium containing 2 μM Venus–Rep–S-pA and incubated for 1 h at 37 °C with rotation. Then, the cells were washed twice with PBS at 37 °C to remove the unbound proteins. Then, the obtained cells were reacted with 2 μM S-pT–Rep–ALOD4 for 1 h at 37 °C with rotation and washed twice with PBS to remove the unbound proteins. The cells resuspended in serum-free DMEM were added to a 3.5 cm glass-bottom dish. The fluorescence signals derived from the S-pA–Venus–Rep-treated cells on the dish were detected using CLSM (FV-300).

Statistical analysis

Data are presented as the means ± standard deviation (SD) of at least three independent trials. Significant differences between treatment means were assessed by one-way ANOVA followed by the Tukey–Kramer multiple comparison test using KaleidaGraph software version 4.1 (Synergy Software). Statistical significance was set at p < 0.05.

Results and discussion

Rep–ALOD4 construction

Rep–ALOD4 (molecular weight: 28 100) comprises Rep fused with ALOD4 via a (GGGS)₂ linker (Fig. 2a). Rep–ALOD4 expressed in E. coli KRX strain, which was transformed with pET–His–Rep–ALOD4, was purified using Ni-affinity chromatography. Unimodal bands appeared around 30 kDa by SDS-PAGE and 28 kDa by size-exclusion chromatography (SEC), indicating Rep–ALOD4 purification (Fig. 2b and ESI Fig. S2a†). The ssDNA-binding ability of Rep–ALOD4 was evaluated by a gel shift assay using agarose gel electrophoresis. S-1 ssDNA with Rep-recognizing sequence (AAGTATTAC) at 5′ end was prepared and reacted with Rep–ALOD4 for arbitrary time or on arbitrary ratio between ssDNA and Rep–ALOD4 (Table 1). By adding S-1 ssDNA to Rep–ALOD4, the band assigned to Rep–ALOD4 shifted toward a higher molecular weight region, suggesting that Rep–ALOD4 conjugated with ssDNA (S-1–Rep–ALOD4) (Fig. 3a). The area of bands corresponding to S-1–Rep–ALOD4 saturated at a 1 : 1 concentration ratio of Rep–ALOD4 to S-1, whereas the reactivity was slightly decreased at a 1 : 2 and 1 : 5 concentration ratio. The reactivity of Rep–ALOD4 with S-1, calculated from the band area, was 62.0% at a 1 : 1 concentration ratio. Furthermore, the reactivity at a 1 : 1 concentration ratio showed a negligible increase for 30 min (Fig. 3b). This trend in Rep–ssDNA reactivity was also observed in other Rep fusion proteins that we have constructed.³⁴ Additionally, Lovendahl et al. reported that ssDNA bound to Rep was released from Rep by adding excesses of ssDNA containing the AAGTATT sequence.³¹ Thus, the conjugation of ssDNA and Rep–ALOD4 was performed at a Rep–ALOD4 to ssDNA concentration ratio of 1 : 1 in subsequent experiments. Since isolating ssDNA-conjugated Rep–ALOD4 from the reaction solution was difficult, we used a mixture of ssDNA-conjugated and unmodified Rep–ALOD4 in subsequent experiments. On the other hand, the ssDNA-binding of Rep is proceeded in the presence of Mg²⁺.³¹ The reactivity of Rep–ALOD4 with S-1 was significantly suppressed by treating EDTA to chelate Mg²⁺, indicating that ssDNA-binding to Rep–ALOD4 was dependent on the ability of Rep (Fig. 3b).

Fig. 2 (a) Rep–ALOD4 structure predicted by AlphaFold3. (b) SDS-PAGE pattern of Rep–ALOD4.

Fig. 3 (a) SDS-PAGE analysis of Rep–ALOD4 reacted with ssDNA at various molar ratios (0, 0.5, 1, 2, and 5 molar equivalents to Rep–ALOD4) for 10 min. (b) The effect of reaction time or 3 mM EDTA on ssDNA conjugation of Rep–ALOD4 (reaction time: 1, 2, 3, 5, 15, and 30 min). The reactivity of Rep–ALOD4 with ssDNA was calculated from the image using ImageJ. Sequence of ssDNA is described in Table 1.

Cholesterol-binding ability of Rep–ALOD4

Next, we verified the cholesterol-binding ability of Rep–ALOD4 using QCM. The lipid bilayer containing cholesterol on the SiO₂-coated sensor chip was prepared by treating liposomes formed from cholesterol and DOPC.^36,37 The ΔF decreased with the treatment of liposomes on the SiO₂-coated sensor chip, suggesting the modification of lipids containing cholesterol and DOPC on the sensor chip (ESI Fig. S3a†). Rep–ALOD4 was treated with the liposome-modified sensor chip to measure the absorption of Rep–ALOD4 with cholesterol. The sensor chip without cholesterol (DOPC : cholesterol = 10 : 0) showed a negligible change in the ΔF with Rep–ALOD4 treatment (Fig. 4a). In contrast, the ΔF of the sensor chip with 20% (w/w) cholesterol (DOPC : cholesterol = 8 : 2) treated with Rep–ALOD4 decreased, and it further decreased with increasing cholesterol content on the sensor chip. Gay et al. reported that ALOD4 recognizes cholesterol in liposomes with >25–30 mol% cholesterol.²⁷ The results for cholesterol recognition of Rep–ALOD4 were in concomitance with those of previous reports.

Fig. 4 (a) Representative QCM profiles for the interaction of Rep–ALOD4 with the SiO₂ sensor modified with liposomes in various ratios of DOPC : cholesterol = 10 : 1, 9 : 1, 8 : 2, 7 : 3, and 6 : 4 (0, 10, 20, 30, 40% w/w cholesterol, respectively). (b) Representative QCM profiles with various concentrations of Rep–ALOD4 treated to the sensor chip modified with liposomes (DOPC : cholesterol = 10 : 0 and 7 : 3). (c) Change in ΔF of 3 μM Rep–ALOD4-treated sensor chip with liposomes (DOPC : cholesterol = 10 : 0 and 7 : 3) by 1 mM Me-β-CD treatment. (d) Representative QCM profiles with Rep–ALOD4- or modified Rep–ALOD4-treated sensor chip with liposomes (DOPC : cholesterol = 7 : 3).

The ΔF of liposomes with 30% (w/w) cholesterol decreased after treating with Rep–ALOD4 in a concentration-dependent manner (Fig. 4b). The plot of the concentration of treated ALOD4 to Δm (μg cm⁻²) was drawn to calculate the association constant of Rep–ALOD4 with cholesterol.^38–40 (ESI Fig. S3b and c†). The experimental adsorption data for the Rep–ALOD4 concentration were linear and the association constant of Rep–ALOD4 with cholesterol was 12.6 × 10⁵ M⁻¹. The association constant of cholesterol with methylated β-cyclodextrins (Me-β-CDs), which has an inclusion complexation with cholesterol, was 1.7 × 10⁴ M⁻¹.⁴¹ Therefore, the interaction between Rep–ALOD4 and cholesterol in liposomes was equivalent to that of Me-β-CDs with cholesterol. To further evaluate the cholesterol-binding ability of Rep–ALOD4, Rep–ALOD4-bound lipid membranes via cholesterol moieties were treated with Me-β-CDs to remove cholesterol from the lipid membrane (Fig. 4c). The lipid membrane containing 40% (w/w) cholesterol was treated with Rep–ALOD4, followed by injecting 1 mM Me-β-CDs. ΔF decreased by the binding of Rep–ALOD4 was gradually increased by injecting Me-β-CDs time-dependently. After 300 min of Me-β-CD injection, the ΔF recovered to the value before binding of Rep–ALOD4 to the lipid membrane, indicating that Rep–ALOD4 was completely detached from the lipid membrane. Nonspecific adsorption of Rep–ALOD4 to lipid membranes would be negligible since Rep–ALOD4 was insignificantly adsorbed on lipid membranes without cholesterol (Fig. 4a). The binding of Rep–ALOD4 to lipid membranes was reflected in the cholesterol content.

The hexapeptide sequence of ALOD4 (GTTLYP; amino acid 98–103) plays an important role in the cholesterol-binding ability of ALOD4.²⁹ To evaluate the interaction between cholesterol and Rep–ALOD4, the GTTLYP sequence of Rep–ALOD4 was substituted with AAAAAA by point mutagenesis (modified Rep–ALOD4) (ESI Fig. S2a and b†). From QCM analysis of the lipid membrane (30% (w/w) cholesterol), the ΔF value of modified Rep–ALOD4 injection was lower than that of Rep–ALOD4 injection (Fig. 4d). This suggests that the binding of Rep–ALOD4 to membrane cholesterol is dependent on the cholesterol-binding ability derived from ALOD4.

The interaction between the plasma membrane and ssDNA-conjugated Rep–ALOD4

Next, the modification of ssDNA on the plasma membrane by ssDNA-conjugated Rep–ALOD4 was evaluated using FITC-modified ssDNA (S-FITC)-conjugated Rep–ALOD4 (S-FITC–Rep–ALOD4) (Table 1). HeLa cells were treated with S-FITC–Rep–ALOD4 in PBS for 10 min and observed by CLSM. A green signal derived from S-FITC–Rep–ALOD4 was detected along the HeLa cell membrane (Fig. 5a). Plasma membrane staining with a lipophilic cyanine dye (DiI) revealed that the green signals derived from S-FITC–Rep–ALOD4 co-localized with DiI-derived signals, indicating that Rep–ALOD4 bound to the plasma membrane (ESI Fig. S4†). HeLa cells treated with 3 and 5 μM S-FITC–Rep–ALOD4 were almost completely wrapped with the green signals, whereas the green signals of cells treated with 1 μM S-FITC–Rep–ALOD4 were scattered on the plasma membrane. As evaluated by flow cytometry, the fluorescence intensity of S-FITC–Rep–ALOD4-treated HeLa cells increased with increasing concentrations of S-FITC–Rep–ALOD4 (Fig. 5b). Moreover, cholesterol-dependent binding of S-FITC–Rep–ALOD4 to the plasma membrane was measured by treatment with Me-β-CDs to extract cholesterol from the plasma membrane. HeLa cells treated with Me-β-CD showed a negligible green signal from S-FITC–Rep–ALOD4 on the plasma membrane and a decreased fluorescence intensity by flow cytometry (Fig. 5a and c). These results suggest that S–FITC–Rep–ALOD4 binds to the plasma membrane with negligible levels of nonspecific adsorption to cells.

Fig. 5 (a) CLSM images of HeLa cells treated with 0, 1, 3, and 5 μM S-FITC–Rep–ALOD4 after 10 min of incubation. In addition, HeLa cells pre-incubated with 1 mM Me-β-CDs for 1 h were treated with S-FITC–Rep–ALOD4. Scale bars: 50 μm. (b) Mean fluorescence intensity of HeLa cells treated with 0, 0.5, 1, 3, and 5 μM S-FITC–Rep–ALOD4 after 10 min of incubation. Me-β-CD means that HeLa cells were pre-treated with 1 mM Me-β-CDs for 1 h at 37 °C. (c) Mean fluorescence intensity of TIG-3, SKOPV-3, MCF-7, and RAW264.7 cells treated with 3 μM S-FITC–Rep–ALOD4 after 10 min of incubation. Data expressed as mean ± SD (n = 3; ****p < 0.0001; NS: not significant).

Furthermore, the binding of S-FITC–Rep–ALOD4 to different cell types such as human normal diploid fibroblast TIG-3, human ovarian cancer SKOV-3, human breast cancer MCF-7, and mouse macrophage-like RAW264.7 cells, was investigated using a flow cytometer (Fig. 5c). TIG-3 and SKOV-3 showed higher fluorescence intensity than HeLa, MCF-7, and RAW264.7. In particular, MCF-7 and RAW264.7 cells had relatively lower fluorescence intensity. It has been reported that ALOD4 adsorbed to the lipid membrane via accessible cholesterol which is a cholesterol without the interaction with sphingomyelin.^42,43 Further detailed evaluation, such as quantification of accessible cholesterol content, is needed to clarify the binding of Rep–ALOD4 to cells.

Next, the time course of the fluorescence signals derived from S-FITC–Rep–ALOD4 on the plasma membrane of HeLa cells was observed in the presence of 10% serum. S-FITC–Rep–ALOD4-derived green signals were partially lost after 30 min of incubation compared to those after 0 min of incubation (Fig. 6a). Subsequent incubation for up to 180 min showed an insignificant change in S-FITC–Rep–ALOD4 signals in HeLa cells. For quantitative analysis, the fluorescence intensity of S-FITC–Rep–ALOD4 in HeLa cells was calculated using image processing. Because the S-FITC–Rep–ALOD4-treated cells adhered to substrates with the incubation, trypsin treatment was necessary to collect incubated cells for subsequent experiments including flow cytometry. However, trypsin-induced digestion could result in the loss of S-FITC–Rep–ALOD4 bound to the plasma membrane. Therefore, the fluorescence intensities were calculated using image processing instead of flow cytometry. The fluorescence intensity of HeLa cells treated with S-FITC–Rep–ALOD4 slightly changed from 30 min to 180 min, whereas that of HeLa cells significantly decreased after 30 min of incubation (Fig. 6b). From these results, S-FITC–Rep–ALOD4 bound to the plasma membrane might be replaced or degraded by contact with serum proteins. The effect of serum proteins on S-FITC and S-FITC–Rep–ALOD4 was evaluated by gel electrophoresis (ESI Fig. S5†). S-FITC and S-FITC–Rep–ALOD4 were incubated with 10% FBS for 30, 60, and 90 min. The fluorescence band assigned to S-FITC insignificantly changed in band intensity by the incubation with 10% FBS (Fig. S5a†). In contrast, the intensity of the band assigned to FITC–Rep–ALOD4 decreased to approximately 50% after 90 min of incubation in the presence of 10% FBS (Fig. S5b†). The decrease rate was concomitant with the fluorescence intensity change of S-FITC–Rep–ALOD4 bound to HeLa cells (Fig. 6b). It suggests that S-FITC–Rep–ALOD4 was partially degraded by serum proteins.

Fig. 6 (a) CLSM images of 3 μM S-FITC–Rep–ALOD4-tethered HeLa cells after 0, 30, 60, 90, 120, and 180 min of incubation in 10% serum-containing medium. Scale bars: 20 μm. (b) Time course for the fluorescence intensities of S-FITC–Rep–ALOD4-tethered HeLa cells, calculated from fluorescence intensity per cell analyzed by ImageJ. Data expressed as mean ± SD of 30 cells. (c) Z-Stack images of HeLa cells treated with 3 μM S-FITC–Rep–ALOD4 for 10 min of incubation. Scale bar: 10 μm.

The distribution of S–FITC–Rep–ALOD4 was scattered on the plasma membrane after incubation in the presence of 10% FBS (Fig. 6a). To further clarify the distribution of S–FITC–Rep–ALOD4 on the plasma membrane, Z-stack images for S-FITC–Rep–ALOD4-treated HeLa cells were acquired (Fig. 6c). Interestingly, S-FITC–Rep–ALOD4-derived green signals were heterogeneously distributed on the plasma membrane. ALOD4 adsorbed to the plasma membrane via non-esterified cholesterol defined as accessible cholesterol.⁴³ Ogasawara et al. reported that the cholesterol in the plasma membrane was classified into two types: sphingomyelin-associated cholesterol and sphingomyelin-free cholesterol.⁴⁴ Sphingomyelin and cholesterol-rich domains in the plasma membrane are called lipid rafts, while sphingomyelin-free cholesterol is defined as accessible cholesterol.^42,43 Thus, S-FITC–Rep–ALOD4 might preferentially adsorb to not lipid raft but other domains in the plasma membrane. Further investigation is required for detailed adsorption distribution and mechanism of Rep–ALOD4.

The signals derived from S-FITC–Rep–ALOD4 were not detected in the cells within 180 min of incubation, indicating that S-FITC–Rep–ALOD4 bound to the plasma membrane was not translocated into the intracellular compartments via endocytosis. The degradation of S-FITC–Rep–ALOD4 by lysosomal protease and DNase could be evaded. Meanwhile, ALOD4 bound to the plasma membrane inhibits low-density lipoprotein-mediated endocytosis and cholesterol transport from the plasma membrane to the endoplasmic reticulum.^28,29,43 This suggests that inhibiting cholesterol-mediated endocytosis by the ALOD4 domain suppresses S-FITC–Rep–ALOD4 internalization into cells. Therefore, we investigated the effect of Rep–ALOD4 on HeLa cell viability. HeLa cells treated with Rep–ALOD4 showed an insignificant increase in LDH release concentration-dependently, indicating that Rep–ALOD4 is less cytotoxic when bound to membrane cholesterol (Fig. S6†). These results indicate that S-FITC–Rep–ALOD4 binding to the plasma membrane was stable for at least 180 min without cytotoxicity in the presence of 10% FBS.

Functional protein displaying on the plasma membrane via complementary base pairing

Rep–ALOD4 tethers Rep-modified ssDNA on the plasma membrane via cholesterol moieties. To clarify the potential applications of Rep–ALOD4, we demonstrated the binding of functional proteins on the plasma membrane via complementary base pairing between ssDNA-conjugated Rep–ALOD4 and the Rep–fusion protein Venus–Rep (Fig. 7a and b). Venus–Rep exhibited the fluorescence at the maximum wavelength of 530 nm (λ_ex = 515 nm), and ssDNA conjugation with Rep-recognition sequence in the same manner as Rep–ALOD4 (Fig. 7c). S-pT and S-pA, model ssDNA sequences with complementary base pairing, were conjugated to Venus–Rep with 84.2 and 81.0% reactivities, respectively (Fig. 7b and Table 1). In contrast, S-pT and S-pA conjugated to Rep–ALOD4 showed 66.7 and 60.0% reactivities, respectively (ESI Fig. S7†).

Fig. 7 (a) Structure of Venus–Rep predicted by AlphaFold3. (b) SDS-PAGE patterns of Venus–Rep and ssDNA-reacted Venus–Rep. Sequence of ssDNA (S-pA and S-pT) described in Table 1. (c) Absorbance and fluorescence spectra of Venus–Rep. λ_ex: 515 nm.

For binding Venus–Rep on the plasma membrane, HeLa cells were first treated with S-pT–Rep–ALOD4 and then with Venus–Rep–S-pA or Venus–Rep–S-pT (Fig. 8a). For instance, Venus binds to the plasma membrane by conjugating S-pT with Rep–ALOD4 (S-pT–Rep–ALOD4) and S-pA with Venus–Rep (Venus–Rep–S-pA), leading to complementary base pairing between plasma membrane-bound S-pT–Rep–ALOD4 and Venus–Rep–S-pA. Complementary base pairs of S-pA and S-pT have been demonstrated to regulate cell–cell interactions by inserting each to the cell surface.⁴⁵ Because the steric hindrance was much greater for cell–cell interactions than for protein–protein interactions, the steric hindrance between S-pT–Rep–ALOD4 and Venus–Rep–S-pA would be negligible on the complementary base pairing.

Fig. 8 (a) Schematic illustration of Venus–Rep tethering on Rep–ALOD4-tethered HeLa cells via the complementary base pairing between S-pT and S-pA. (b) CLSM images of HeLa cells treated with 6 μM S-pT–Rep–ALOD4, followed by 6 μM Venus–Rep, Venus–Rep–S-pA or Venus–Rep–S-pT. Scale bars: 20 μm. (c) Fluorescence histogram and (d) mean fluorescence intensities of HeLa cells treated with S-pT–Rep–ALOD4, followed by Venus–Rep–S-pA or Venus–Rep–S-pT. Data are expressed as means ± SD (n = 3; ****p < 0.0001; NS, not significant).

The CLSM images showed that the treatment of unmodified Venus–Rep had insignificant yellow signals, derived from Venus, on unmodified and S-pT–Rep–ALOD4-treated HeLa cells (Fig. 8b). The non-specific interaction of Venus–Rep with the plasma membrane and S-pT–Rep–ALOD4 was negligible. The treatment of Venus–Rep–S-pT to S-pT–Rep–ALOD4-treated HeLa cells showed weak yellow signals on the plasma membrane. Meanwhile, the combination of Venus–Rep–S-pA with S-pT–Rep–ALOD4 in HeLa cells showed significant fluorescence signals along the plasma membrane, suggesting the display of Venus–Rep–S-pA on the plasma membrane via complementary base pairing. Flow cytometry revealed that the S-pT–Rep–ALOD4-treated HeLa cells exhibited significantly higher fluorescence intensity with Venus–Rep–S-pA than with Venus–Rep–S-pT (Fig. 8c and d). Since treating S-pT–Rep–ALOD4-treated HeLa cells with Venus–Rep–S-pT increased the fluorescence intensity, Venus–Rep might be internalized into the cells by endocytosis. This indicated that complementary base pairing between plasma membrane-bound S-pT–Rep–ALOD4 and Venus–Rep–S-pA contributed to Venus–Rep tethering on the plasma membrane. Consequently, S-pT–Rep–ALOD4 was conjugated to the plasma membrane based on the cholesterol-binding ability of ALOD4, and Venus–Rep–S-pA was bound via complementary base pairing between S-pT and S-pA. Rep–ALOD4 has potential applications as an ssDNA-tethering molecule on the plasma membrane based on the interaction between membrane cholesterol and ALOD4.

Conclusions

In this study, we developed Rep–ALOD4, which can bind ssDNA and membrane cholesterol as an ssDNA-tethering molecule on the plasma membrane of cells. The display of ssDNA on the plasma membrane by Rep–ALOD4 did not require the chemical modification of ssDNA with a hydrophobic anchor, only the addition of the Rep-recognizing sequence (AAGTATTAC) to arbitrary DNA at 5′ end. Furthermore, functional proteins such as Venus–Rep were conjugated to the plasma membrane modified with Rep–ALOD4 via complementary base pairing between S-pT and S-pA. However, S-FITC–Rep–ALOD4 heterogeneously distributed on the plasma membrane via membrane cholesterol. The effect of the heterogeneous distribution of ssDNA–Rep–ALOD4 on the subsequent function of ssDNA on the plasma membrane is unknown. Further investigation is required for detailed distribution and mechanism of Rep–ALOD4 binding to the plasma membrane. ssDNA-binding Rep–ALOD4 may be useful for tethering DNA on the plasma membrane of cells via cholesterol. Taken together, Rep–ALOD4 will be a valuable material for tethering ssDNA on the plasma membrane via cholesterol, which has various potential applications such as nano-assembly formation, drug delivery, and diagnostics.

Author contributions

Conceptualization of the research was conceived by K. N., E. K., and M. M. Formal analysis and investigation were performed by M. I. and K. N. Visualization and writing of the manuscript were done by K. N. Reviewing and editing of the manuscript was performed by all authors. Funding acquisition was provided by K. N., E. K. and M. M. This work was completed under the supervision of E. K. and M. M.

Data availability

All data regarding this manuscript are already presented in the graphs of the main paper and the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the Japan Society for the Promotion of Science (JSPS, grant number 23K17209 to K. N., 23K28427 to M. M.) and the Terumo Life Science Foundation (M. M.). This work was the result of using the research equipment shared in the MEXT Project for promoting public utilization of advanced research infrastructure (program for supporting the construction of core facilities; grant number JPMXS0440200021).

References

1. H. Yang, L. Yao, Y. Wang, G. Chen and H. Chen, Chem. Sci., 2023, 14, 13325–13345.
2. E. S. Yolcu, N. Askenasy, N. P. Singh, S. E. L. Cherradi and H. Shirwan, Immunity, 2002, 17, 795–808.
3. K. Keren, Eur. Biophys. J., 2011, 40, 1013–1027.
4. A. Gubu, X. Zhang, A. Lu, B. Zhang, Y. Ma and G. Zhang, Mol. Ther. – Nucleic Acids, 2023, 33, 144–163.
5. Z. Li, J. Wang, M. P. O'Hagan, F. Huang, F. Xia and I. Willner, ACS Nano, 2023, 17, 15308–15327.
6. N. C. Seeman and H. F. Sleiman, Nat. Rev. Mater., 2018, 3, 17068.
7. S. Chen, Z. Xu, S. Li, H. Liang, C. Zhang, Z. Wang, J. Li, J. Li and H. Yang, Angew. Chem., 2022, 134, 1–6.
8. D. Sarkar, P. K. Vemula, G. S. L. Teo, D. Spelke, R. Karnik, L. Y. Wee and J. M. Karp, Bioconjugate Chem., 2008, 19, 2105–2109.
9. M. Xiong, G. Kong, Q. Liu, L. Liu, Y. Yin, Y. Liu, H. Yuan, X. B. Zhang and W. Tan, Nano Lett., 2023, 23, 183–191.
10. J. Wang, J. Song, X. Zhang, S. M. Wang, B. Kang, X. L. Li, H. Y. Chen and J. J. Xu, J. Am. Chem. Soc., 2023, 145, 1273–1284.
11. Y. Hu, Y. Duan and K. Salaita, Angew. Chem., Int. Ed., 2023, 62(30), e202302967.
12. B. Zhao, C. O’Brien, A. P. K. K. Karunanayake Mudiyanselage, N. Li, Y. Bagheri, R. Wu, Y. Sun and M. You, J. Am. Chem. Soc., 2017, 139, 18182–18185.
13. G. Feng, X. Luo, X. Lu, S. Xie, L. Deng, W. Kang, F. He, J. Zhang, C. Lei, B. Lin, Y. Huang, Z. Nie and S. Yao, Angew. Chem., 2019, 131, 6662–6666.
14. R. Ueki, S. Atsuta, A. Ueki and S. Sando, J. Am. Chem. Soc., 2017, 139, 6554–6557.
15. D. Y. Lee, B. H. Cha, M. Jung, A. S. Kim, D. A. Bull and Y. W. Won, J. Biol. Eng., 2018, 12, 1–11.
16. A. Schoenit, E. A. Cavalcanti-Adam and K. Göpfrich, Trends Biotechnol., 2021, 39, 1208–1220.
17. X. Xiong, H. Liu, Z. Zhao, M. B. Altman, D. Lopez-Colon, C. J. Yang, L. Chang, C. Liu and W. Tan, Angew. Chem., 2013, 125, 1512–1516.
18. S. F. Jones, H. Joshi, S. J. Terry, J. R. Burns, A. Aksimentiev, U. S. Eggert and S. Howorka, J. Am. Chem. Soc., 2021, 143, 8305–8313.
19. M. Anaya, M. Kwak, A. J. Musser, K. Müllen and A. Herrmann, Chem. – Eur. J., 2010, 16, 12852–12859.
20. C. Gosse, A. Boutorine, I. Aujard, M. Chami, A. Kononov, E. Cogné-Laage, J. F. Allemand, J. Li and L. Jullien, J. Phys. Chem. B, 2004, 108, 6485–6497.
21. Y. Bagheri, S. Chedid, F. Shafiei, B. Zhao and M. You, Chem. Sci., 2019, 10, 11030–11040.
22. Y. Li, X. Chen, C. Lv and Y. Cheng, Front. Chem., 2023, 11, 1–10.
23. Y. Teramura, Y. Kaneda, T. Totani and H. Iwata, Biomaterials, 2008, 29, 1345–1355.
24. H. Liu, B. Kwong and D. J. Irvine, Angew. Chem., Int. Ed., 2011, 50, 7052–7055.
25. J. G. Shannon, C. L. Ross, T. M. Koehler and R. F. Rest, Infect. Immun., 2003, 71, 3183–3189.
26. R. W. Bourdeau, E. Malito, A. Chenal, B. L. Bishop, M. W. Musch, M. L. Villereal, E. B. Chang, E. M. Mosser, R. F. Rest and W. J. Tang, J. Biol. Chem., 2009, 284, 14645–14656.
27. A. Gay, D. Rye and A. Radhakrishnan, Biophys. J., 2015, 108, 1459–1469.
28. R. S. Chakrabarti, S. A. Ingham, J. Kozlitina, A. Gay, J. C. Cohen, A. Radhakrishnan and H. H. Hobbs, eLife, 2017, 6, 1–27.
29. R. E. Infante and A. Radhakrishnan, eLife, 2017, 6, 1–23.
30. Y. Mashimo, H. Maeda, M. Mie and E. Kobatake, Bioconjugate Chem., 2012, 23, 1349–1355.
31. K. N. Lovendahl, A. N. Hayward and W. R. Gordon, J. Am. Chem. Soc., 2017, 139, 7030–7035.
32. W. Guo, Y. Mashimo, E. Kobatake and M. Mie, Nanotechnology, 2020, 31, 255102.
33. M. Mie, T. Niimi, Y. Mashimo and E. Kobatake, Biotechnol. Lett., 2019, 41, 357–362.
34. K. Nishida, G. Wang, E. Kobatake and M. Mie, ACS Biomater. Sci. Eng., 2023, 9, 5260–5269.
35. M. Mie, R. Hirashima, Y. Mashimo and E. Kobatake, Appl. Sci., 2020, 10, 1–10.
36. L. Bar, M. E. Villanueva, S. Neupane, G. Cordoyiannis and P. Losada-Pérez, Phys. Status Solidi A, 2023, 220(22), 2200625.
37. H. Tae, C. Yang and N. J. Cho, Acc. Mater. Res., 2022, 3, 1272–1284.
38. I. Lundstrom, Prog. Colloid Polym. Sci., 1985, 70, 76–82.
39. R. A. Latour, J. Biomed. Mater. Res., Part A, 2015, 103, 949–958.
40. M. Tanaka, A. Mochizuki, T. Motomura, K. Shimura, M. Onishi and Y. Okahata, Colloids Surf., A, 2001, 193, 145–152.
41. H. W. Frijlink, A. C. Eissens, N. R. Hefting, K. Poelstra, C. F. Lerk and D. K. F. Meijer, Pharm. Res. An Off. J. Am. Assoc. Pharm. Sci., 1991, 8, 9–16.
42. J. Ando, M. Kinoshita, J. Cui, H. Yamakoshi, K. Dodo, K. Fujita, M. Murata and M. Sodeoka, Proc. Natl. Acad. Sci. U. S. A., 2015, 112, 4558–4563.
43. K. A. Johnson, S. Endapally, D. C. Vazquez, R. E. Infante and A. Radhakrishnan, J. Biol. Chem., 2019, 294, 17289–17300.
44. F. Ogasawara, F. Kano, M. Murata, Y. Kimura, N. Kioka and K. Ueda, Sci. Rep., 2019, 9, 1–10.
45. X. Li, X. Xie, Z. Ma, Q. Li, L. Liu, X. Hu, C. Liu, B. Li, H. Wang, N. Chen, C. Fan and H. Song, Adv. Mater., 2018, 30, 1–8.

---

Footnotes

† Electronic supplementary information (ESI) available: Complete experimental method and Fig. S1–S7. See DOI: https://doi.org/10.1039/d4bm01127a

‡ Present affiliation: Department of Life Science and Technology, School of Life Science and Technology, Institute of Science Tokyo, 4259 Nagatsuta-cho, Midori-ku, Yokohama, 226-8501, Japan.

---